The explosive growth in Internet use during the last few years has made it possible to communicate with greater numbers of people on-line than with any other technology. Recent software advances have made it possible to transmit and collect secure data from remote individuals over the Internet in an efficient and interactive manner. Thus, the internet has become an extraordinarily powerful potential resource for performing questionnaire-based research. One of the most enticing, yet unexplored, medical applications of the Internet is the possibility of performing clinical trials entirely on-line. Because of the vast scope of the Internet, it may be possible to study the attributes of various compounds which might otherwise never be evaluable in traditional clinic- based settings. For example, there exist a number of nutritional compounds which may be modestly effective in relieving osteoarthritis symptoms. Because of the large numbers and prohibitive costs involved in detecting efficacy from these compounds, it is unlikely that they will all be adequately evaluated in traditional clinical trials. The Internet, on the other hand, using validated symptom questionnaires, could have great utility in testing these safe compounds in the treatment of osteoarthritis. On the other hand, we need to learn about many factors which are critical to the feasibility of Internet-based trials, such as response rates, demographic characteristics of respondents, willingness of respondents to participate, validity of responses, protocol compliance, and participant retention. Our aim, therefore, is to evaluate aspects of the feasibility, utility and validity of performing a clinical trial using the Internet, by performing a model on-line trial of a glucosamine/chondroitin sulfate nutritional supplement among individuals with symptomatic knee osteoarthritis.